The archive operation should only archive open messages

Dirk Hohndel hohndel at infradead.org
Thu Apr 15 14:07:37 PDT 2010


On Thu, 15 Apr 2010 13:41:17 -0700, Carl Worth <cworth at cworth.org> wrote:
> One of the searches that I use most frequently, (for mail that I want to
> respond to on a fairly timely basis), is
> 
> 	tag:inbox and tag:to-me		[*]
> 
> Sometimes, this search will show a large mailing-list thread with only a
> few messages open. Perhaps part-way through the thread, someone started
> to CC me. Or perhaps my address got dropped from the CC at some
> point. Either way, I am presented with a subset of the messages from the
> thread, even though all of the thread's messages are in my inbox still.
> 
> That much is just fine. I'm giving priority to messages where people
> thought I would be particularly interested, and that's just as it should
> be.
> 
> A bad bug occurs when paging through the thread with the space
> bar. After showing me these few messages, it will then proceed to
> archive *all* the messages in the thread (not only those it showed
> me). And I'm likely to be unaware of this since the closed (but not yet
> archived) messages are not easily distinguished from messages that were
> previously closed and archived.
> 
> Some people will claim (and I've even agreed) that the space bar is too
> magic. But this bug also happens with an explicit command to archive the
> current thread (such as hitting 'a').
> 
> I think the fix is to change these commands to only archive the messages
> that are currently open. That will make these operations behave as I
> expect, and I don't think will cause any unexpected or confusing
> behavior. But please let me know if you disagree.

I am always confused about the behavior of 'a' - does it archive the
current message? Or the current thread? Or the current thread down to
where I am? Or (as you propose) just the open messages?

I think we really need to spend some time to crsiply define the
semantics of these commands.

/D

-- 
Dirk Hohndel
Intel Open Source Technology Center


More information about the notmuch mailing list